This resubmission for a second competitive renewal project in our program of research to develop and test treatments for late-life generalized anxiety disorder (GAD) is consistent with PA 07-163, which supports research to reduce the burden of mental illness in older adults and increase efficiency and accessibility of treatment. The original study (MH53932, 01 - 05) demonstrated the efficacy of cognitive behavior therapy (CBT), relative to a minimal contact control condition (MCC), in an academic clinical trial (Stanley et al., 2003a). The ongoing renewal project (R01 MH53932, 06 - 09) extends this research to primary care. Preliminary analyses suggest greater improvement in worry and anxiety severity in CBT than UC, with improved outcomes at 6-months relative to 3-months. Attrition in CBT is low. Effect sizes are moderate and comparable to results from prior collaborative care studies. In the proposed second renewal project (years 10 - 14), we will further promote the generalizability to contemporary real-world settings by examining the effectiveness of using Anxiety Clinic Specialists (ACS) and non-expert Counselors to deliver the intervention, relative to usual care (UC). Several enhancements to our intervention support its sustainable implementation in routine clinical practice: (1) Increased collaboration with primary care through the use of electronic medical records (EMR) to identify patients and communicate with providers;(2) CBT modifications that involve modular treatment, integration of telephone-based service delivery, incorporation of patient preference, briefer sessions and simplified procedures, more attention to coexistent depression and motivation/life values, and extended duration of care using telephone-based booster sessions;and (3) Implementation in two diverse medical settings (insurance-based primary care, Veterans Affairs Medical Center) to enhance generality of findings. A total of 222 medical patients, age 60 and over, with GAD will be recruited through primary care clinics at the new study sites. Patients will be assigned randomly to CBT-ACS (expert), CBT-Counselor (non-expert), or Usual Care. Treatment will last 6 months. Pilot data support our ability to train non-experts and recruit a more diverse and severe patient sample. Outcomes are generally equivalent to or greater than the ongoing trial. Patients who complete 6 months of CBT will enter a follow-up phase, with evaluations at 9, 12, and 18 months. Patients who exit UC after 6 months will be offered CBT. Specific Aims will compare outcomes in CBT-ACS, CBT-Counselor, and UC over 6 months. Primary outcomes address severity of worry, anxiety, and GAD. Secondary outcomes assess coexistent symptoms (depression, sleep, pain, alcohol use), functional status, and satisfaction. Long-term maintenance of gains will be examined over the subsequent 12 months in the two CBT groups. Exploratory Aims will examine resource use/cost and potential moderators and mediators of outcome.